Mains-driven zero-crossing detector uses only a few high-voltage parts
You can get isolation with low power consumption using this simple circuit.
Luca Matteini, Agliana, Italy; Edited by Paul Rako and Fran Granville -- EDN, December 1, 2011
The circuit in this Design Idea generates a zero-crossing pulse off the ac mains and provides galvanic isolation. The falling edge of the output pulse happens at approximately 200 μsec before the zero crossing. You can use the circuit to safely stop the triggering of a thyristor gate, giving it time to properly turn off. The circuit generates short pulses only when the mains voltage is approximately 0V, thereby dissipating only 200 mW at 230V and a 50-Hz input.
The circuit charges capacitor C1 up
to the limit that 22V zener diode D3
creates (Figure 1 and Reference 1).
You limit the input current with resistors
R1 and R5. As the input-rectified
voltage drops below the C1 voltage, Q1
starts conducting and generates a pulse
a few hundreds of microseconds long.
The coupling of IC1 makes the response
of Q1 squarer. The rms operating voltage
dictates the only requirement for R1
and R5. SMD, 1206-size resistors typically
withstand 200V-rms operation.
This design splits the input voltage
between R1 and R5, for a total rating of
400V rms. D3 limits the voltage across
the bridge to 22V so that all of the subsequent components can have lower
voltage ratings. A 22V zener diode can
clamp as high as 30V, so this design
uses a 50V, 470-nF ceramic capacitor.
Ceramic capacitors have better reliability
than electrolytic or tantalum capacitors, especially at higher temperatures.
If you prefer a cheaper and
smaller 25V part, you can change the
zener diode’s voltage to 18V and still
have a good margin for safety. Use R4 to
limit the peak current in the LED. The
primary limit on the LED current is
the slope of the rectified ac input. The
gradual slope doesn’t let Q1 generate
current spikes when it discharges C1’s
stored energy.
This design splits the input voltage
between R1 and R5, for a total rating of
400V rms. D3 limits the voltage across
the bridge to 22V so that all of the subsequent components can have lower
voltage ratings. A 22V zener diode can
clamp as high as 30V, so this design
uses a 50V, 470-nF ceramic capacitor.
Ceramic capacitors have better reliability
than electrolytic or tantalum capacitors, especially at higher temperatures.
If you prefer a cheaper and
smaller 25V part, you can change the
zener diode’s voltage to 18V and still
have a good margin for safety. Use R4 to
limit the peak current in the LED. The
primary limit on the LED current is
the slope of the rectified ac input. The
gradual slope doesn’t let Q1 generate
current spikes when it discharges C1’s
stored energy.
You can simulate the operation
of the circuit in LTspice Version IV
(Figure 2 and Reference 2). With a
230V input at 50 Hz, the simulation
shows a 17-mA peak in the optocoupler LED. The simulation
gives good results
with inputs of 90 to 250V,
both at 50 and 60 Hz. At
110V and a 60-Hz input,
the LED current peak is 8.5
mA, so IC1 still works. If
you need higher LED-drive
currents, you can reduce
the value of R3 or increase
the value of C1.Testing a physical circuit shows good correlation with the simulation (Figure 3). Driving the isolated output from a 5V logic supply yields a good pulse waveform (Trace 1). The mains input is fed to the scope with a 15V isolation transformer for safety (Trace 2). You can use the persistence feature of the oscilloscope to show the zero-crossing point when zooming in to the transition (Figure 4). This approach allows you to accurately measure the pulse timing relative to the input zero crossing.

|
References |
|
Talkback
-
Good idea, but looks like there is a misinterpretation of withstand voltage of 1206 resistors. The withstand voltage is voltage that could be applied to resistor during 1 minute and resistor should survive. In this case we must talk only about maximum working voltage that could be at least 2X less than withstand voltage.
So, selection of 200V withstand voltage 1206 resistors for this application is definitely not correct and 60% of maximum working voltage for resistors must be used for a reliable design.
Vladimir Doubovis - 2012-10-2 19:31:29 PST -
The single resistor alternative solution works fine as well, provided you limit current to a "safe" value: enough for the limits imposed by CTR to obtain a sharp response, and even a bit more than that to compensate LED aging (hence CTR derate). In a circuit where LED is on most of the time, that can be an issue, in 24/7 operation.
Working with 115 Vac is simpler than with 230 Vca, for power dissipation. With a resistor-only limiter and the former rating (even if not avoiding LED aging effects) circuit should go fine with a power requirement nearly the double of the proposed circuit.
L Matteini - 2011-12-12 06:07:30 PST -
I achieve much the same effect by simply connecting a resistor in series with the LEDs of an AC input optoisolator across the AC line. That's 2 parts, or 3 if you need a pullup on the output transistor. Total cost, maybe 30 cents.
Mike Eck - 2011-11-12 17:39:48 PST -
I think that I can see a use for this circuit as the reset for a ramp generator in a triac-controlled voltage regulator circuit
William Ketel - 2011-5-12 14:40:07 PST






















